Metrics¶

All URIs are relative to https://api.ubiops.com/v2.1

metrics_create¶

MetricDetail metrics_create(project_name, data)

Create metrics

Description¶

Create a custom metric. The name must start with custom..

Required Parameters¶

• name: Name of the metric
• metric_type: Type of the metric. It can be either 'delta' or 'gauge'.

Optional Parameters¶

• description: Description of the metric
• unit: Unit of the metric
• labels: A list of labels that can be used to get data points containing the metric. For example, if the metric is defined for a deployment version and will be queried later with the ID of the deployment version, the labels list should contain 'deployment_version_id'.

Request Examples¶

{
  "name": "custom.metric-1",
  "metric_type": "delta",
  "unit": "seconds",
  "labels": ["deployment_version_id"]
}

Response Structure¶

Details of the created metric

• id: Unique identifier for the metric
• name: Name of the metric
• description: Description of the metric
• creation_date: The date when the metric was created
• last_updated: The date when the metric was last updated
• custom: A boolean indicating whether the metric is custom
• metric_type: Type of the metric
• unit: Unit of the metric
• labels: A list of labels that can be used to get data points containing the metric

Response Examples¶

{
  "id": 10,
  "name": "custom.metric-1",
  "description": "My custom metric",
  "creation_date": "2023-09-01T08:32:14.876451Z",
  "last_updated": "2023-09-01T10:52:23.124784Z",
  "custom": true,
  "metric_type": "delta",
  "unit": "seconds",
  "labels": ["deployment_version_id"]
}

Example¶

Initialize core_api using your credentials.

project_name = 'project_name_example' # str
data = ubiops.MetricCreate() # MetricCreate

# Create metrics
api_response = core_api.metrics_create(project_name, data)
print(api_response)

Parameters¶

Return type¶

MetricDetail

Authorization¶

API token

[Back to top]

metrics_delete¶

metrics_delete(project_name, metric_name)

Delete metric

Description¶

Delete a metric. Only custom metrics can be deleted.

Example¶

Initialize core_api using your credentials.

project_name = 'project_name_example' # str
metric_name = 'metric_name_example' # str

# Delete metric
core_api.metrics_delete(project_name, metric_name)

Parameters¶

Return type¶

void (empty response body)

Authorization¶

API token

[Back to top]

metrics_get¶

MetricDetail metrics_get(project_name, metric_name)

Get metric

Description¶

Retrieve details of a metric

Response Structure¶

Details of a metric

• id: Unique identifier for the metric
• name: Name of the metric
• description: Description of the metric
• creation_date: The date when the metric was created
• last_updated: The date when the metric was last updated
• custom: A boolean indicating whether the metric is custom
• metric_type: Type of the metric
• unit: Unit of the metric
• labels: A list of labels that can be used to get data points containing the metric

Response Examples¶

{
  "id": 10,
  "name": "custom.metric-1",
  "description": "My custom metric",
  "creation_date": "2023-09-01T08:32:14.876451Z",
  "last_updated": "2023-09-01T10:52:23.124784Z",
  "custom": true,
  "metric_type": "delta",
  "unit": "seconds",
  "labels": ["deployment_version_id"]
}

Example¶

Initialize core_api using your credentials.

project_name = 'project_name_example' # str
metric_name = 'metric_name_example' # str

# Get metric
api_response = core_api.metrics_get(project_name, metric_name)
print(api_response)

Parameters¶

Return type¶

MetricDetail

Authorization¶

API token

[Back to top]

metrics_list¶

list[MetricDetail] metrics_list(project_name, custom=custom)

List metrics

Description¶

List available metrics in the project

Optional Parameters¶

• custom: A boolean indicating whether to list default or custom metrics for the project

Response Structure¶

A list of details of metrics

• id: Unique identifier for the metric
• name: Name of the metric
• description: Description of the metric
• creation_date: The date when the metric was created
• last_updated: The date when the metric was last updated
• custom: A boolean indicating whether the metric is custom
• metric_type: Type of the metric
• unit: Unit of the metric
• labels: A list of labels that can be used to get data points containing the metric

Response Examples¶

[
  {
    "id": 1,
    "name": "deployments.requests",
    "description": "Requests to a deployment version",
    "creation_date": "2023-09-01T08:32:14.876451Z",
    "last_updated": "2023-09-01T10:52:23.124784Z",
    "custom": false,
    "metric_type": "delta",
    "unit": "requests",
    "labels": ["deployment_version_id", "user_id"]
  },
  {
    "id": 2,
    "name": "deployments.credits",
    "description": "Credits usage",
    "creation_date": "2023-09-02T10:12:51.195381Z",
    "last_updated": "2023-09-02T10:12:51.195381Z",
    "custom": false,
    "metric_type": "delta",
    "unit": "credits",
    "labels": ["deployment_version_id"]
  },
]

Example¶

Initialize core_api using your credentials.

project_name = 'project_name_example' # str
custom = True # bool (optional)

# List metrics
api_response = core_api.metrics_list(project_name, custom=custom)
print(api_response)

Parameters¶

Return type¶

list[MetricDetail]

Authorization¶

API token

[Back to top]

metrics_update¶

MetricDetail metrics_update(project_name, metric_name, data)

Update metric

Description¶

Update a metric. Only custom metrics can be updated.

Optional Parameters¶

• name: Name of the metric
• description: Description of the metric
• unit: Unit of the metric
• labels: A list of labels that can be used to get data points containing the metric. For example, if the metric is defined for a deployment version and will be queried later with the ID of the deployment version, the labels list should contain 'deployment_version_id'.

Request Examples¶

{
  "name": "custom.metric-2"
}

Response Structure¶

Details of the updated metric

• id: Unique identifier for the metric
• name: Name of the metric
• description: Description of the metric
• creation_date: The date when the metric was created
• last_updated: The date when the metric was last updated
• custom: A boolean indicating whether the metric is custom
• metric_type: Type of the metric
• unit: Unit of the metric
• labels: A list of labels that can be used to get data points containing the metric

Response Examples¶

{
  "id": 10,
  "name": "custom.metric-2",
  "description": "My custom metric",
  "creation_date": "2023-09-01T08:32:14.876451Z",
  "last_updated": "2023-09-01T10:52:23.124784Z",
  "custom": true,
  "metric_type": "delta",
  "unit": "seconds",
  "labels": ["deployment_version_id"]
}

Example¶

Initialize core_api using your credentials.

project_name = 'project_name_example' # str
metric_name = 'metric_name_example' # str
data = ubiops.MetricUpdate() # MetricUpdate

# Update metric
api_response = core_api.metrics_update(project_name, metric_name, data)
print(api_response)

Parameters¶

Return type¶

MetricDetail

Authorization¶

API token

[Back to top]

time_series_data_aggregate¶

object time_series_data_aggregate(project_name, data)

Aggregate metric data

Description¶

Aggregate metric data for given date, metrics and labels. Only data up to 2 minutes ago is accepted.

Required Parameters¶

• metric: Name of the metric
• labels: Dictionary containing key/value pairs where key indicates the string that can be used to query this metric later and value is the corresponding value of that
• data: A list of dictionaries containing 'date' and 'value' fields to indicate the value of the metric for a specific date

Request Examples¶

[
  {
    "metric": "deployments.requests",
    "labels": {
      "deployment_version_id": "056efa9e-67eb-45e3-a49a-0742b3f08aee"
    },
    "data": [
      {
        "date": "2023-09-15T20:12:33.210+00:00",
        "value": 182
      },
      {
        "date": "2023-09-15T21:41:12.532+00:00",
        "value": 1
      }
    ]
  }
]

Example¶

Initialize core_api using your credentials.

project_name = 'project_name_example' # str
data = ubiops.TimeSeriesDataCreate() # TimeSeriesDataCreate

# Aggregate metric data
api_response = core_api.time_series_data_aggregate(project_name, data)
print(api_response)

Parameters¶

Return type¶

object

Authorization¶

API token

[Back to top]

time_series_data_create¶

object time_series_data_create(project_name, data)

Create metric data

Description¶

Insert data points for a metric. Multiple metrics for different types is supported.

Required Parameters¶

• metric: Name of the metric
• labels: Dictionary containing key/value pairs where key indicates the string that can be used to query this metric later and value is the corresponding value of that
• data: A list of dictionaries containing 'date' and 'value' fields to indicate the value of the metric for a specific date. The value is inserted for the minute provided in the date field.

Request Examples¶

[
  {
    "metric": "deployments.requests",
    "labels": {
      "deployment_version_id": "056efa9e-67eb-45e3-a49a-0742b3f08aee"
    },
    "data": [
      {
        "date": "2023-09-15T20:00:00.000+00:00",
        "value": 182
      },
      {
        "date": "2023-09-15T21:00:00.000+00:00",
        "value": 1
      }
    ]
  }
]

Example¶

Initialize core_api using your credentials.

project_name = 'project_name_example' # str
data = ubiops.TimeSeriesDataCreate() # TimeSeriesDataCreate

# Create metric data
api_response = core_api.time_series_data_create(project_name, data)
print(api_response)

Parameters¶

Return type¶

object

Authorization¶

API token

[Back to top]

time_series_data_list¶

TimeSeriesDataList time_series_data_list(project_name, metric=metric, start_date=start_date, end_date=end_date, aggregation_period=aggregation_period, labels=labels, unit_period=unit_period)

List time series data

Description¶

List data points for a metric

Available metrics for deployments:

• deployments.requests: Number of requests to a deployment version
• deployments.failed_requests: Number of failed requests to a deployment version
• deployments.request_duration: Average time in seconds for a deployment request to complete
• deployments.input_volume: Volume of incoming data in bytes
• deployments.output_volume: Volume of outgoing data in bytes
• deployments.network_in: Volume of inbound data traffic in bytes
• deployments.network_out: Volume of outbound data traffic in bytes
• deployments.queue_size: Average number of queued requests
• deployments.queue_time: Average time in seconds for a request to start processing
• deployments.memory_utilization: Average memory used during a request
• deployments.instances: Number of active deployment instances
• deployments.credits: Usage of credits, calculated by multiplying the credit rate of a deployment instance type by the number of hours the deployments are running

Available metrics for pipelines:

• pipelines.requests: Number of requests to a pipeline version
• pipelines.failed_requests: Number of failed requests to a pipeline version
• pipelines.request_duration: Average time in seconds for a pipeline request to complete
• pipelines.input_volume: Volume of incoming data in bytes
• pipelines.output_volume: Volume of outgoing data in bytes
• pipelines.object_requests: Number of object requests in a pipeline version
• pipelines.object_failed_requests: Number of failed object requests in a pipeline version

Required Parameters¶

• metric: Name of the metric
• start_date: Start date for metric data points
• end_date: End date for metric data points

Optional Parameters¶

• aggregation_period: Time period in seconds in which data points are grouped. It defaults to the highest resolution possible given the provided date range. Available values are: 60, 300, 900, 3600, 7200, 21600 and 86400. Start and end dates are adjusted according to the aggregation period. For example, if aggregation period is 3600, start date is rounded down to the previous full hour and end date is rounded up to the next full hour.
• labels: Comma-separated values for labels to filter on data points. It must be in the format: key-1:value-1,key-2:value-2.

Request Examples¶

With aggregation period 60, to get the credits usage of a deployment version per minute

{
  "metric": "deployments.credits",
  "start_date": "2024-05-01T10:00:00Z",
  "end_date": "2024-05-01T10:30:12Z",
  "labels": "deployment_version_id:dbcc9de3-1dcb-48ad-8197-3b2ac99f5e94"
}

With aggregation period 3600, to get the credits usage of a deployment version per hour

{
  "metric": "deployments.credits",
  "start_date": "2024-05-01T10:00:00Z",
  "end_date": "2024-05-01T16:30:12Z",
  "labels": "deployment_version_id:dbcc9de3-1dcb-48ad-8197-3b2ac99f5e94"
}

Response Structure¶

• metric: Name of the metric
• metric_type: Type of the metric
• unit: Unit of the metric
• start_date: Start date for metric data points
• end_date: End date for metric data points
• aggregation_period: Time period in seconds in which data points are grouped
• labels: Labels to filter on data points
• data: A list of dictionaries containing the data points

Response Examples¶

{
  "metric": "deployments.requests",
  "metric_type": "delta",
  "unit": "requests/s",
  "start_date": "2023-01-01T10:00:00Z",
  "end_date": 2023-01-01T12:00:00Z",
  "aggregation_period": 3600,
  "labels": {
    "deployment_version_id": "8935a589-8686-4ce7-8c9e-8b5e529c6b47"
  },
  "data": [
    {
      "start_date": "2023-01-01T10:00:00Z",
      "end_date": 2023-01-01T11:00:00Z",
      "value": 3
    },
    {
      "start_date": "2023-01-01T11:00:00Z",
      "end_date": 2023-01-01T12:00:00Z",
      "value": 10
    }
  ]
}

Example¶

Initialize core_api using your credentials.

project_name = 'project_name_example' # str
metric = 'metric_example' # str (optional)
start_date = 'start_date_example' # str (optional)
end_date = 'end_date_example' # str (optional)
aggregation_period = 56 # int (optional)
labels = "label1:value1,label2:value2" # str (optional)
unit_period = 56 # int (optional)

# List time series data
api_response = core_api.time_series_data_list(project_name, metric=metric, start_date=start_date, end_date=end_date, aggregation_period=aggregation_period, labels=labels, unit_period=unit_period)
print(api_response)

Parameters¶

Return type¶

TimeSeriesDataList

Authorization¶

API token

[Back to top]

time_series_delete¶

time_series_delete(project_name, time_series_id)

Delete time series

Description¶

Delete a time series

Example¶

Initialize core_api using your credentials.

project_name = 'project_name_example' # str
time_series_id = 'time_series_id_example' # str

# Delete time series
core_api.time_series_delete(project_name, time_series_id)

Parameters¶

Return type¶

void (empty response body)

Authorization¶

API token

[Back to top]

time_series_search¶

list[TimeSeriesSearch] time_series_search(project_name, metric=metric, labels=labels, custom=custom, exact_match=exact_match, limit=limit, offset=offset)

Search time series

Description¶

Search through time series

Optional Parameters¶

• metric: Name of the metric
• labels: Comma-separated values for labels to filter on data points. It must be in the format: key-1:value-1,key-2:value-2.
• custom: A boolean indicating whether only default or custom metrics should be returned. If this parameter is not provided, both types are returned.
• exact_match: A boolean indicating whether the provided labels should match exactly or whether matching a subset is allowed. Defaults to false (matching a subset is allowed).
• limit: The maximum number of time series to return. It defaults to 500.
• offset: The number that indicates the starting point of the time series to return. It defaults to 0.

Response Structure¶

A list of time series

• metric: Name of the metric
• labels: Labels that the time series has
• resolution: Metric resolution in seconds

Response Examples¶

[
  {
    "metric": "deployments.requests",
    "labels": {
      "deployment_version_id": "8935a589-8686-4ce7-8c9e-8b5e529c6b47",
      "user_id": "5bb50513-2b4e-466a-ab88-e5be70d63f75"
    },
    "resolution": 60
  }
]

Example¶

Initialize core_api using your credentials.

project_name = 'project_name_example' # str
metric = 'metric_example' # str (optional)
labels = "label1:value1,label2:value2" # str (optional)
custom = True # bool (optional)
exact_match = True # bool (optional)
limit = 56 # int (optional)
offset = 56 # int (optional)

# Search time series
api_response = core_api.time_series_search(project_name, metric=metric, labels=labels, custom=custom, exact_match=exact_match, limit=limit, offset=offset)
print(api_response)

Parameters¶

Return type¶

list[TimeSeriesSearch]

Authorization¶

API token

[Back to top]